I have a summarized dataset in the below format to track the performance of an email campaign. Each of the flag columns have a binary value , and the total count of customers corresponding to that row is given in Count column.   EDIT : So, the first row signifies that there are 10 customers which were in Control set, who didn't open the email , didn't respond and were not funded. The second row signifies 20 customers in Test set who didn't open email, didn't respond and were not funded. End of EDIT    Campaign Control_Flag Email_Open Responder Funded Count $Revenue 1 Y 0 0 0 10   1 N 0 0 0 20   1 N 1 0 0 30   1 Y 0 1 0 40   1 N 0 1 0 50   1 N 1 1 0 60   1 Y 0 1 1 70 $ 700,000 1 N 0 1 1 80 $ 800,000 1 N 1 1 1 90 $ 900,000     I need to prepare the report in the below format from it. For Responders/Funded/Revenue rows, I just need the sum of counts where these flag values are 1. Total Test is the sum of Email Open and Email Not Open columns.     Campaign 1   Test     Email Not Open Email Open Total Test Control Target 150 180 330 120 Responders 130 150 280 110 Funded 80 90 170 70 Revenue $   800,000 $             900,000 $ 1,700,000 $ 700,000     Running the below code show the counts of both responders and non responders. How do I show counts of only those cases where the flag variable is 1?   proc report data=TARGETS2_G; column resp control_flag,cust; define resp / group; define control_flag / across; define cust /analysis sum; run;   EDIT 2 : Uploaded the input file. ... View more